MOTM

On Tue, 23 May 2000 davevosh@... wrote:

> crow,
> i`d love to see a source of uncorrelated ( say 3 or 4 ) random triggers 
> derived from such a noise source. any chance ?
> best,
> dave v.

  Uncorrelated?  That would require good old shot noise to be truly random
between concurrent triggers.  In a PIC, I can rig several long feedback
shift registers and thus create pseudo-random noise by instructing the
regs to shift at a rate of 40KHz, but unless the seed number can be set by
some sort of "real-world random" for each register, the triggers would
actually be correlated by the finite (long, yes, but still finite) shift
register bit patterns.

  Hm, now this has me thinking.  I have methods I've used in the past to
get fairly "true random" seeds that involve detecting the error in
thermocouple cold-junction compensation servo loops, but that might not be
the best solution here (as I'm not trying to build a process controller
here). What might be of more use to us however is some way of setting
specific random seeds by hand and then load/start the registers.

  A "random" sequencer for triggers is what I'm arriving at here.  Some
way to set a seed with a load jack/pushbutton, a clock rate switch to run
it at 40KHz for audio uses or select a manual shift clock from a VCO or
whatever.  I'll have to puzzle over it a bit to see if what I'm thinking
actually works the way I want it to.

  And here I just wanted to use my PIC part that I use to replace the
flimsy digital noise chip Prophet-5s used.  Now look what it has become.
;)

Crow

/**/